JMIR Publications

ABSTRACT

Background:

Artificial intelligence (AI)-enabled digital mental health modalities are increasingly being explored for mental health digital therapeutics, such as symptom reduction and treatment personalization. However, a comprehensive synthesis of the literature is lacking on how AI-enabled digital mental health modalities are used for mental health management and how they facilitate access to mental health care for key end users in both formal and informal settings.

Objective:

This scoping review explored and mapped the available literature on how AI-enabled digital mental health modalities are used for mental health management and mental health access by mental health care providers, the general public, and clinical populations across formal and informal settings.

Methods:

The scoping review was guided by the Levac et al. (2010) methodology. Primary studies written in English and published between 2020 and 2025 were retrieved from PubMed, Embase, Scopus, CINAHL, and IEEE Xplore databases. Included data sources were required to involve adults aged 18 years or older from: (1) the general public; (2) clinical populations with any mental health condition, or (3) healthcare providers either using AI-enabled digital mental health modalities for their own mental health or delivering such modalities to their patients. The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) Extension for Scoping Reviews guided narrative synthesis and reporting of findings.

Results:

Out of 2073 screened studies, 50 met the inclusion criteria and were included. The studies were predominantly conducted in the United States, followed by Canada. Most studies explored AI models and text-based conversational agents utilized by clinical populations and the general public. Narrative synthesis of findings generated two overarching themes: (1) AI-Enabled Pathways for Mental Health Management Optimization and (2) AI-enabled Mental Health Care Access Pathways. Theme 1 included sub-themes: (a) Proactive Mental Well-Being, (b) Predictive Analytics, (c) Improvement in Skills and Symptomology, and (d) Maintenance and Monitoring. Theme 2 included sub-themes: (a) Accessibility, (b) Acceptability, (c) Accommodation, and (d) Affordability. AI-enabled digital mental health modalities were used to support mental health management through symptom identification and reduction, predicting risks and treatment outcomes, monitoring symptomology, and delivering psychoeducation. Many of the AI-enabled digital mental health modalities were regarded as acceptable by users. Fewer studies reported on the affordability and accessibility of the included modalities.

Conclusions:

AI-enabled digital mental health modalities are becoming ubiquitous in mental health care management. However, the availability of many of the included modalities does not mean they foster access to mental health care. More research addressing how these modalities are used to improve access to mental health care is required. Additional longitudinal research is necessary to inform understanding of how these modalities can be implemented directly by mental health care providers in formal settings and beyond research trial phases.